Unless my interviews go better than I expect them to I'm working on a prober that automatically updates probes for data systems where the read path can be fairly heterogenous like SQL where the result size and the query time can vary fairly widely.